Why all this fuss about CTRL(X) ??

[Hall]

In article <184 at devon.UUCP> paul at devon.UUCP (Paul Sutcliffe Jr.) writes:
>In article <1955 at watdragon.UUCP>, jmsellens at watdragon.UUCP (John M. Sellens) writes:
>> Make a file called ctrl.h with lines like
>> #define CTRLA	'\001'
>> in it (should take mere seconds).  Put #include "ctrl.h" in your
>> C source files and use CTRLA.  No fuss, no muss.  Easy to understand.
>> Portable. And it even works.
>
>That's fine if you know ahead of time what control characters you 
>want to play with.  However, if I want to find the ``control'' value
>of a variable, as in:
>	..[example]..
>then your method, although good for what it was intended, won't do.

But then, neither did the initial definition for CTRL(X) that started
this whole mess.

Recall the initial definition:

#define CTRL(X) ('X' ^ 0100)

or something very similar that works only for the Reiser cpp (& similar).

John Sellens' idea for cntrl.h leads to yet another possibility that maybe
would implement CTRL for either cpp (Reiser or ANSI) in a way that would
be compatible with the original definition (i.e. no 's required).

Suppose that you have a file ctrl.h that contains:

#define __CTRLA '\001'
#define __CTRLB '\002'
...
#define __CTRLZ '\003'

#define CTRL(X) __CTRL/**/X	/* Reiser cpp */
#define CTRL(X) __CTRL//**/**/X	/* (May need to be this, to preserve X */

#define CTRL(X) __CTRL # X	/* ANSI cpp */

This would still have some problems with things like CTRL(@) and CTRL(?),
but it comes close..

Marcus Hall
..!ihnp4!ihlpl!marcus